Compare models
Pick up to 4models. Specs render side-by-side. Share the URL — it's stateless.
 Mistral Large 2Mistral |  Gemini 3 Pro |  Llama 4 ScoutMeta |  DeepSeek-V3.1DeepSeek | |
|---|---|---|---|---|
| Vendor | Mistral | Meta | DeepSeek | |
| Family | Mistral | Gemini | Llama | DeepSeek |
| Release date | 2024-07-24 | 2026-04-22 | 2025-04-05 | 2025-08-21 |
| Context window | 128,000 tokens | 2,000,000 tokens | 10,000,000 tokens | 128,000 tokens |
| Parameters | 123B | — | 109B | 671B |
| Modality | text | text, vision, audio, video | text, vision | text |
| License | Mistral Research License | proprietary | Llama 4 Community License | MIT |
| Source | open weights | proprietary | open weights | open weights |
| Description | — | Google DeepMind's flagship multimodal model. Best-in-class for multimodal understanding and agentic / vibe coding workflows. A "Deep Think" reasoning mode is available to AI Ultra subscribers. | — | Large MoE open-weight model. Predecessor to DeepSeek-V4. |
| Links |
Add a model
Gemini 2.5 Flashmax
Claude 4.6 Sonnetmax
GPT-5.5maxDeepSeek-V4-Flashmax
Qwen3.6-27BmaxGemini 3 Pro✓DeepSeek-V4-PromaxClaude 4.7 OpusmaxClaude 4.5 HaikumaxDeepSeek-V3.1✓GPT-5max
Grok 4maxQwen3 235Bmaxo3maxGPT-4.1maxLlama 4 Scout✓Llama 4 MaverickmaxGemini 2.5 PromaxDeepSeek-R1maxMistral Large 2✓Max 4 models. Remove one to add another.